一种改进的彩色遥感图像边缘检测算法研究

An Improved Algorithm For Edge Detection of Color Remote Sensing Images

针对当前方法在对彩色遥感图像做边缘检测时,普遍存在无法精确定位边缘、边缘信息获取不完整以及受到噪声影响等问题,提出一种改进的边缘检测算法。采用分数阶微分和Canny算子相结合的方法,设计一种分数阶微分模板应用到Canny算子中,同时引入高斯曲率滤波理论,在对彩色遥感图像做边缘检测的同时平滑图像,寻找最小正则化能量点,获得算法的最优参数,从而得到更优的彩色遥感图像边缘检测效果。经过对比实验后发现,改进的算法在对彩色遥感图像边缘检测的过程中,能够准确定位边缘获得更多的边缘信息,平滑图像噪声以及抑制噪声的非线性放大。

There are some generally problems such as the inability to accurately locate the edges,the incomplete acquisition of edge information and the influence of noise at the current method in the edge detection of color remote sensing images.To solve these problems,an improved edge detection algorithm is proposed.Using a combination of fractional differential and Canny operator,a fractional differential template was applied to the Canny operator.At the same time,the Gaussian curvature filtering theory was introduced to smooth the image,find the minimum regularized energy point and obtain the optimal parameters of the algorithm while performing edge detection on color remote sensing images.Thereby,better edge detection effect of color remote sensing image was obtained.The comparison experiments show that the improved algorithm can accurately locate the edge to obtain more edge information,smooth image noise and suppress nonlinear amplification of noise in the process of edge detection of color remote sensing images.